Generalized Likelihood Ratio Test for GNSS Spoofing Detection in Devices with IMU

Spoofing attacks in global navigation satellite systems (GNSSs) aim at inducing the estimation of a fake position at the victim receiver. Many devices, including smartphones, are nowadays equipped with both a GNSS receiver and an inertial measurement unit (IMU), which also provides location/movement information, while being immune from GNSS attacks. We propose a spoofing detection technique based on the comparison between GNSS and IMU measurements. The detection is performed through a generalized likelihood ratio test (GLRT), which is efficiently implemented by a matrix multiplication approach. In particular, the device a) estimates its orientation from magnetometer and gyroscope measurements, b) estimates its position, acceleration, and velocity by a maximum likelihood approach, and c) performs the GLRT for spoofing detection. The performance of the proposed GLRT is compared with the Kalman filter innovation test and with the direct comparison method (DCM), both in terms of false alarm/missed detection probabilities and computational complexity.